What would a clean-sheet set of wiring regulations look like today?

It is an interesting thought experiment to consider how things might have been done differently if it wasn't for historical accident..

Years ago we weren't that fussed about ADS, just made sure that everything was solidly earthed. If we'd continued down that line we might now have c.p.c.s that a 4x the c.s.a. of the line conductors (so even during earth faults exposed parts remain below 50V) and probably concentric earth cables (perhaps think SWA but with copper armour strands) so far less reliance on RCDs etc.

Part of the idea behind ring circuits was instead of having multiple final circuits, you run the distribution circuit to the socket and have a final circuit for each appliance (far better discrimination, and better use of resources as copper is shared between different uses, making the best use of diversity). Ideal when you want lots of sockets for relatively low power appliances. Had we carried on in that vein, especially with the vast increase in the number of appliances, we might have ended up with domestics based on something more like flush dado bus-bar trunking with sockets just clipped in where ever they're wanted, no lighting circuits, just a FCU in each room for the lights, and possibly high demand appliances like cookers and showers tapping into the same system. So no CU as such, just one or two big fuses (maybe one per floor).

  - Andy.

I've thought quite a bit about this over the last few days.

It'd be a very interesting way to do things, although a CPC with a quarter the linear resistance of the phase conductors could get very expensive.

If all copper we'd end up with 6mm2 cooker circuits with a total copper area of ~36mm2, which is getting towards some feeder cables!
I suppose they'd be a big drive in favour of three phase circuits in that regard, because replacing a 45A single phase with 16A three-phase circuit would reduce the copper area to 4x1.5+6 = 12mm2. Time for three phase radial socket circuits with different sockets on different phases?

It would be even worse for truly chunky whole-floor rings I would think. 16mm2 of phase conductor area would have a total copper area of 96mm2!.

I suppose you could try something like the American "MC-AP" cable from Southwire. That uses copper phase/neutral conductors with an aluminium earthing conductor which is in intimate contact with an interlocking aluminium armour sheath.

Hard to get concrete figures, but a 10AWG (~5.3mm2) copper single phase cable has an 8AWG (8.4mm2) continuous aluminium conductor in contact with aluminium armour that is something like 26mm2 of cross section. So total of 34mm2 of CPC.

Even adjusting for aluminium vs copper and the armour being interlocking rather than a welded tube,  that probably gets close to what you suggest! Just a thought.

Probably easier to work with than pyro, which would achieve the 4x CSA requirement more or less by default.

Well the old lead sheathed paper was a  bit like that - a massive outer covering compared to the cores. 
Like Wensleydale, I do like a bit of pyro but not unlike SWA, the mechanical skills to install it well are in short supply. 

Given we can do ADS, and the RCD has been invented, luckily we don't need it any more, I do like the idea of a distributed CU, with fuses at the light switches or lighter loads.

Personally I'd also be up for 3 phase sockets, no need to keep phases in separate rooms. There seem to be no great excess of phase to phase accidents and these 16A 3 phase things are moderately common in domestic Dutch kitchens for cookers and larger instant water heaters

I think we tend to put 3 phase on a pedestal of being in some way harder and more dangerous to do than single, yet in terms of difficulty,  in many ways 4 thin cores fused at a lower level, are easier than 2 fatter ones backed by a higher PSSC.
regards Mike.
PS some other Dutch kitchens wire the same style of connector as 2 ganged single phases so a meter is useful.

If you wired a chunky single phase super-ring in 10mm, you could put something like a 4-pole 55A RCBO on it (Reference Method A, D curve MCB and time delay RCD possibly?), although I perhaps see some benefit in armoured cable given the consequences of cutting into it! 

With reasonable spreading of loads around the ring that's going to have a real load capability of something like 75A, basically no risk of thermal overload and not too much copper use. If you put a couple of thin data cores in the ring you could also do DALI or similar and avoid all the lighting complexity - connect switch or lamp however convenient and the data cables from the super rings all common back in the vestigial consumer unit.

You could do it all in 3-phase, although an efficient ring would probably want an 8-pole RCBO... or more likely a four pole RCD connected to two four pole MCB. Presumably in some sort of compact package. 10mm is probably excessive for a three phase submain, 6mm would still handle 38A each way around (Ref Method A) on three-phase.

Branch circuits could have a 30mA RCBO at the branch off point. I could imagine a socket faceplate with 4 IEC 60906-1 sockets and a separate RCBO all in a standard back box - kind of like those RCD sockets that were popular for retrofit for a while.

Fuses for things that are unlkely to be overloaded. With separate trips on each side of the ring we can also connect cooker-sized point loads wherever is convenient without risking overloads.

Comparatively cheap many-pole devices seem to open all sorts of opportunities!

Would you need that heavy a cable ? - a typical UK 80 or 100 A supply for the whole house is similar total KVA to 3phases at 32A each, and that could  be done in a 4mm radial, (or even a ring of 2,5mm)

You could also save copper & energy by encouraging  voltages of 690/400 and/or  1k3/690 to augment  the current 230/400 for those places where high loads and long runs mean voltage drops are an issue.

M

PS the only country using 60906-1  as written is South Africa. The Brazilian variants use 4mm (10A) or 4.8mm(20A) pins instead of 4.5mm(16A), and the Swiss ones have a different offset of the earth pin from the centre line (5mm not 3mm). Apart from that its a truly common standard (!)

I guess I'd be a bit worried that a 32A per phase circuit is going to get overloaded by someone connecting big single phase loads like fan heaters to sockets in the wrong combinations. Especially if lazy builders don't select their phases properly.

4mm or 6mm ring would seem to mitigate against that.

I suppose a 2.5mm ring would still handle 20A per side or so, which is probably enough for the floor of a house, assuming all the big loads are three phase inherently. A heavier ring (4mm for 30A/side, 6mm for 38A/side) would probably be a good idea in a hybrid situation where many of the loads are single phase still.

I think 690V is probably better suited to slightly heavier commercial installations, such as restaurants and commercial kitchens. Those big pizza convection ovens would be the obvious example. .

On 60906-1 - I suggested it primarily because I like the idea of packing sockets in, and I'd prefer a socket standard that was compact and designed for a pretty high real-world rating. I think the layman will try to use sockets for their nominal rating continuously, and that you can never have enough sockets.

I think the bulk of gains from standardisation are probably achieved by just getting a socket compatible with Europlug, rather than going all the way to Schuko or CEE 7/5. And since we'd probably want shutters anyway I don't think the Swiss standard would be fully suitable either.

I think you'd struggle to get more than two CEE 7 sized sockets onto a wall plate, compared to four for the 60906-1, especially once we consider the need for an overcurrent device in the faceplate. Probably arrange things so you can chain a second quad off the first one's RCBO if you want, for 8 socket points (fitting the French limit on radial circuits).

If we standardise on 1.5mm XLPE lighting cable it can do that too!.

I guess I'd be a bit worried that a 32A per phase circuit is going to get overloaded by someone connecting big single phase loads like fan heaters to sockets in the wrong combinations. Especially if lazy builders don't select their phases properly.

4mm or 6mm ring would seem to mitigate against that.

I suppose a 2.5mm ring would still handle 20A per side or so, which is probably enough for the floor of a house, assuming all the big loads are three phase inherently. A heavier ring (4mm for 30A/side, 6mm for 38A/side) would probably be a good idea in a hybrid situation where many of the loads are single phase still.

I think 690V is probably better suited to slightly heavier commercial installations, such as restaurants and commercial kitchens. Those big pizza convection ovens would be the obvious example. .

On 60906-1 - I suggested it primarily because I like the idea of packing sockets in, and I'd prefer a socket standard that was compact and designed for a pretty high real-world rating. I think the layman will try to use sockets for their nominal rating continuously, and that you can never have enough sockets.

I think the bulk of gains from standardisation are probably achieved by just getting a socket compatible with Europlug, rather than going all the way to Schuko or CEE 7/5. And since we'd probably want shutters anyway I don't think the Swiss standard would be fully suitable either.

I think you'd struggle to get more than two CEE 7 sized sockets onto a wall plate, compared to four for the 60906-1, especially once we consider the need for an overcurrent device in the faceplate. Probably arrange things so you can chain a second quad off the first one's RCBO if you want, for 8 socket points (fitting the French limit on radial circuits).

If we standardise on 1.5mm XLPE lighting cable it can do that too!.